[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: RE: Water probe: signal processing now ok



Original poster: "Denicolai, Marco" <Marco.Denicolai@xxxxxxxxxxx>

Hi Paul,

> > Inverse convolution in the time domain is very unstable.
>
> The deconvolution is implemented by FFT, division, then
> inverse FFT, so I'm not sure where your instability is coming
> from - it should be the same as your first solution.  Is
> there another method for deconvolution?

Sorry, I meant that I noticed inverse convolution in time domain is very
unstable. Therefore I do it in the frequency domain.
To be more exact, the impulse response calculation is very sensitive to
the noise amount in the FFT. That's why I abandoned the time domain
approach.

> > the secondary is indeed responsible for that [100kHz] peak.
>
> Glad that's confirmed.  We need to remove that peak from the
> impulse response  (we dont' want to deconvolute that
> resonance - it's part of the system being measured!).

How you would do that best? Hand retouching of the FFT?

> What of the resonances at 5.6Mhz and 9.2Mhz?  I believe you
> concluded these were genuine probe responses, not coax resonances?

Yes, these are resonances of the upper arm. I detected the same in my
smaller probe prototype and they are always there whatever the signal
buffering used is (fiber, videoamp, etc.

> > Note that the responses are so clean now because we have been
> > averaging (within the o-scope) 200 readings of the step.
>
> Are you still using the fiber and video amp?  Can you email
> me the latest scope trace data for the step?

I'll send those to your email address.

>
> > It seems to me like the next step is to switch Thor on and to make
> > some measurements at full operation.
>
> Yes, and you'll also need to do some below-breakout firing to
> calibrate models.  It will be a major step forward to have
> some precision topvolts waveform data.  At last there is a
> hope for moving forward on the issue of discharge loading.

How did you think to calibrate more the models? Do you mean using Thor's
simulations as a start to calculate its theoretical output voltage?

Best Regards